Skip to main content

Full text of "Galaxies with unusually high abundances of molecular hydrogen"

See other formats


Galaxies with Unusually High Abundances of Molecular Hydrogen 



O 

o 

(N 



> 
o 
\o 
\o 
a^ 
o 
\o 
o 

o 



A. V. KasparovEQ 
Sternberg Astronomical Institute. 

A. V. Zaso\l!| 

Sternberg Astronomical Institute. 

A sample of 66 galaxies from the catalog of Bettoni et al. (CISM) with anomalously high 
molecular-to-atomic hydrogen mass ratios {Mmoi/MHi > 2) is considered. The sample galaxies 
do not differ systematically from other galaxies in the catalog with the same morphological types, in 
terms of their photometric parameters, rotational velocities, dust contents, or the total mass of gas 
in comparison with galaxies of similar linear sizes and disk angular momentum. This suggests that 
the overabundance of H2 is due to transition of HI — > H2. Galaxies with bars and active nuclei are 
found more frequently among galaxies which have Mmoi estimates in CISM. In a small fraction of 
galaxies, high Mmoi/Mui ratios are caused by the overestimation of Mmoi due to a low conversion 
factor for the translation of CO-line intensities into the number of H2 molecules along the line of 
sight. It is argued that the "molecularization" of the bulk of the gas mass could be due 1) to the 
concentration of gas in the inner regions of the galactic disks, resulting to a high gas pressure and 
2) to relatively low star-formation rate per unit mass of molecular gas which indeed takes place in 
galaxies with high Mmoi/Mni ratios. 

Astronomy Reports, 2006, Vol. 50, No. 8, p. 626-637. 



1. INTRODUCTION 

The densest and coolest component of the inter- 
stellar medium — molecular gas — is observed in 
all types of disk galaxies. In many cases, a mass of 
molecular gas inside galactic disks is comparable 
to a mass of neutral hydrogen {HI). The amount 
and spatial density of molecular gas reflects the 
specific properties of the evolution of the inter- 
stellar medium in galactic disks. The most mas- 
sive gaseous formations — giant molecular clouds, 
which affect the dynamical evolution of the disk, — 
are associated with molecular gas. However, more 
importantly, the mass and density of the molecu- 
lar gas determine the star-formation rate and star- 
formation efficiency in galactic disks. 

Despite the crucial role that molecular gas plays 
in galaxies, the factors determining its abundance 
relative to neutral gas remain poorly known. Most 
information on interstellar molecular gas has been 
obtained from measurements of CO-line intensi- 
ties — CO has a very low excitation temperature. 
The conversion of the CO-line luminosity into the 
total mass of gas is a separate problem. For 
galaxies with normal heavy-element abundances 
the conversion factor connecting the CO-linc in- 
tensity with the number of H2 molecules along the 
line of sight, X = N{H2)/ Ico, is usually set equal 
to (2 — 3) • 10^*^ mol/{K ■ km/s), however, the 
universality of this parameter remains a topic of 
discussion. Adopting a constant X value is, of 
course, a simplification, especially when galaxies 
with strongly differing mass, gas density, or star- 



' akasparovaOyandex.rul 



izasoviasai . msu. rnl 



formation intensity are compared. Various meth- 
ods for measuring X and the dependence of X on 
the galactic luminosity show appreciable scatter, 
together with a systematic increase in X with de- 
creasing luminosity (mass) of the galaxy and de- 
creasing heavy-element abundance (Q, The 
conversion factors for dwarf galaxies and giant lu- 
minous galaxies may differ by an order of magni- 
tude. We also expect X to be smaller for dense cir- 
cumnuclear disks, where the molecular gas seems 
to be contained in both clouds and the diffuse in- 
tercloud medium (Q, H). 

Various authors (0, @, 0) have analyzed the 
conditions for molecular-to- atomic, Mmoi — > Mhi, 
transformations and vice- versa (see also references 
in these papers). The rate at which molecules are 
formed depends on both the density (or pressure) 
of a gas and on the density of UV flux, which de- 
stroys molecules. According to Elmegreen (0), 
the Mmoi/Mni ratio is proportional to P'^ '^/j, 
where P is the gas pressure and j is the UV flux 
density. The gas compression in shocks can also 
play an important role in the "molecularization" 
of hydrogen @ . 

Since the gas pressure depends both on its sur- 
face density and on the volume density of the stel- 
lar disk, it decreases rapidly with galactocentric 
distance. Indeed, the azimuthally averaged den- 
sity ratio Mmoi /Mhi follows the pressure in such 
a way that the contribution of Mmoi to the total 
mass of gas becomes small at a periphery of the 
galaxy 

Available estimates show that the integrated ra- 
tio Mmoi /Mhi is appreciably lower than unity 
for most spiral galaxies. According to Casoli et 
al. 01, Mmoi /Mhi is, on the average, close to 0.2. 
Boselli et al. Jlj found the average molecular-to- 
atomic gas mass ratio to be Mmoi/ Mhi = 0.14 



for a sample of 266 galaxies. However, some spi- 
ral galaxies are known to have anomalously high 
Mmoi/Mni values, exceeding two or even three (an 
order of magnitude higher than the average value) . 
Our aim here is to analyze the specific features of 
such galaxies with a predominance of molecular gas 
and possible factors that could lead to such high 
Mmoi/Mni estimates. 

As our initial data, we used "A new catalogue 
of ISM content of normal galaxies" (CISM) of Bet- 
toni et al. This catalog contains interstellar- 

medium data for almost two thousand "normal" 
galaxies, more than 300 of which have estimates 
for their molecular-gas mass. Normal galaxies are 
considered to be those that display no morpholog- 
ical distorsion (such as tidal tails or bridges) or 
strong perturbations of their circular velocities, al- 
though several galaxies among those included in 
the catalog can, nevertheless, be considered to be 
interacting galaxies (see below). The molecular- 
gas mass estimate is based on the adopted value 
of X = 2.3 ■ lO^o mol/K ■ km/s, taken to be the 
same for all galaxies. 



2. MAIN PROPERTIES OF GALAXIES 
WITH A PREDOMINANCE OF H2 

2.1. The Galaxy Sample 

We selected a total of 66 objects with anoma- 
lously high ratios Mm.ni I Mhj > 2 from the CISM 
catalog of Bettoni et al. [lO|- Forty- five of these 
galaxies have Mmoi/MHi > 3. Table |31 gives 
the main properties of these sample galaxies. We 
adopted data on the interstellar medium from 
CISM; photometric and kinematic properties from 
the HYPERLEDA database, and the pho- 
tometric scale in accordance with the catalog of 
Baggett et al. The columns of the table give: 
a running number for the galaxy, its name, mor- 
phological type (as a number), heliocentric dis- 
tance (in Mpc), radial photomeric scale length (in 
kpc), B-band luminosity (in solar luminosities), 
photometric diameter D25 (in kpc), logarithm of 
the dust mass (in solar units), logarithm of the 
HI mass (in solar units), logarithm of the mass of 
molecular gas (in solar units) , logarithm of the ro- 
tational velocity derived from the widths of HI (in 
km/s), and corrected U-B and 5- F color indices. 
The last column ("Remarks") shows the presence 
of a bar (B), membership in the Virgo or Coma 
clusters (V or C), and signs of interaction (VV 
number according to Vorontsov-Velyaminov) . 

Figure ^ compares the galaxies in our sample 
with the CISM galaxies with known molecular- 
gas contents (below, we will consider only this 
part of all CISM catalog) in terms of their mor- 
phological types, surface brightnesses SB (within 
the isophotal diameter D25) relative to the av- 



TABLE 1 ; Mean values of some parameters for galaxies 
of the CISM catalog (sample 1) and for sample galaxies 
with Mmoi/Mni > 2 (sample 2), with the standard 
deviations (STD) and the statistical probability of the 
absence of differences between the two samples (p). 



Sample 


Mean STD p, % 


Morphological type 


1 


2.80 3.06 95.00 


2 


2.24 2.24 


Magnitude 


1 


-20.75 1.37 99.00 


2 


-21.06 1.16 


logVrot 


1 


2.19 0.21 95.00 


2 


2.23 0.20 


log{Mdust/Mgas) 


1 


-3.38 0.39 99.95 


2 


-3.21 0.36 



erage surface brightness (SB) for CISM galaxies 
of the same type, absolute B-band magnitude M, 
and rotational velocity Vrot derived from the HI 
line widths. These distributions show that galax- 
ies with high Mmoi/Mffj ratios are mostly Sa — 
Sbc galaxies {t = 1 — 4), with virtually no late- 
type galaxies among them. Note that allowing for 
the luminosity dependence of the conversion factor 
does not affect the conclusion that galaxies with 
late morphological types have lower Mmoi/MHi ra- 
tios o. 

The galaxies of our sample are essentially indis- 
tinguishable from the other CISM galaxies in terms 
of the brightness of their stellar disks and their 
luminosities, although the mean value for M is 
shifted somewhat toward higher luminosities. The 
rotational velocities Vrot for the sample galaxies 
and for all the CISM galaxies with available Mmoi 
estimates have more or less the same distributions, 
ranging from 50 to 400 km/s. Table Ogives the 
numerical results of these comparisons: the mean 
values and standard deviations of the quantities 
compared, and the probability that the difference 
between the means of the two samples is of random 
nature. We conclude that the objects in our sam- 
ple do not differ strongly from the other galaxies 
in terms of the parameters described above. 

Note that galaxies with optical bars make up 
a significant fraction of galaxies with measured 
molecular-gas masses (Table Whereas barred 
galaxies account for about one-third of all the 
CISM galaxies, they comprise one-half of all galax- 
ies with available molecular-gas mass estimates 
(which evidently bias to objects with high H2 
abundances). The latter is also true for the galax- 
ies in our sample with Mmoi/Mni > 2. It follows 
from Table |2 that this sample is characterized by 



60 




FIG. 1: Distributions of the galaxies in morphological type t, absolute B-band magnitude M, logarithm of 
the surface density normalized by the mean surface density for the corresponding morphological type, and the 
logarithm of the rotational velocity, Vrot- The light gray histograms show the distributions for all galaxies of 
catalog .10!] with estimated Mmoi values, while the dark gray histograms are for galaxies in our sample, that is a 
predominance of molecular gas. 



high frequency of occurrences of bars and of nu- 
clear activity (the latter was taken from the cata- 
log 01). This may be indirectly associated with 
the concentration of a gas toward galactic centers. 
The existence of a possible correlation between the 
abundance of molecular gas and the level of nuclear 
activity was earlier pointed out in . 

Thus, galaxies with high Mmoi/Mni ratios do 
not differ significantly from other galaxies in their 
photometric properties, morphological properties, 
or rotational velocities. However, the question re- 
mains open as to what extent the abundance of 
atomic and molecular gas in such galaxies is really 
anomalous. 



2.2. HI and H2 Abundances in the Sample 
Galaxies 

A high M„ioi/Mhi ratio could be due to either 
a high abundance of molecular gas or a reduced 
mass HI in the galaxy. Let us now compare the 



TABLE 2: Percentage of barred galaxies and galaxies 
with active nuclei (according to [l4Qfor (1) all objects 
of CISM O, (2) galaxies of CISM|10l with molecular- 
hydrogen mass estimates, and (3) the galaxies of our 
sample with high relative abundances of molecular gas. 



Sample 


Bar 


Active nucleus 


N 


1 


35.3 ±0.8 


7.7 ±0.2 


1916 


2 


53.3 ±3.0 


15.8 ±0.9 


317 


3 


56.7 ±6.9 


22.4 ±2.7 


66 



galaxies in our sample with Mmoi/Mni > 2 and 
all the CISM galaxies of the corresponding mor- 
phological types. The left-hand panel in Fig. [21 
shows the distribution of the sample galaxies in 

lg{o-mol/ {<yraol)), whcrC dynol = Mmol/Dl^ is the 

surface density of the molecular gas within the 
optical radius, and (amoi) is the mean a„ioi for 
all galaxies of the same morphological type with 
known masses Mmoi- The right-hand panel in 



Fig. 121 shows a similar comparison of the sur- 
face density of neutral hydrogen, cthi- The 
mean amoi for the sample galaxies is higher by 
Alg(fTmo// {o'moi)) ~ 0.5 compared to the average 
value for all galaxies of the corresponding type. 
The HI abundance in the sample galaxies is char- 
acterized by an asymmetric shape of the distribu- 
tion of (J HI (note the more gently sloping "wing" 
for galaxies with strong deficiencies of neutral hy- 
drogen), although the mean ani does not differ 
strongly from the value {(J hi) for all galaxies of 
the same morphological type. If we select galax- 
ies with the lowest HI abundances, with surface 
densities cthi less than half the mean value, we 
will find that their mean lg{(T,noi/ {cimoi)) is also 
higher than zero, indicating that, in this case too, 
we are dealing with an excess of molecular gas. 
Thus, galaxies with high Mmoi/Mni ratios really 
possess high H2 abundances, rather than having 
lost their HI as a result of, e.g., its being swept 
out from the galaxy as a more tenuous component 
of the interstellar medium. 



2.3. Total Gas Mass in the Disks of the 
Sample Galaxies 

We now compare the total gas contents of the 
galaxies using the relation between the gas mass 
and the specific angular momentum of the disk, 
which is characterized by the product of the ro- 
tational velocity Vrot and radius R of the disk. 
As the disk radius, we used either the photomet- 
ric (isophotal) radius R25 or the photometric ra- 
dial disk scale Ro (in linear units). The relation 
between the mass of gas and VrotR is obeyed by 
most of the Sb — Ir galaxies with various bright- 
nesses, both single objects and galaxies with close 
companions 0, 0| (see also references therein). 
Although this relation has a physical basis, and 
follows from the fact that the gas surface densities 
in most of the galaxies are (on average) close to 
the threshold density for local Jeans instability, or 
proportional to this density, here we treat it as a 
purely empirical relation. 

Figure|31(left, top and bottom) shows the depen- 
dence of the mass of atomic gas Mhi on the spe- 
cific angular momentum. All set of CISM galaxies 
and galaxies from our sample are shown by filled 
and open circles, respectively. Most of the sam- 
ple galaxies display HI deficits, which are espe- 
cially strong among galaxies with relatively low 
angular momenta. The situation changes radi- 
cally if instead of Mhi we plot the total mass 
of atomic and molecular gas (including helium) 
Mgas = l-4(M/f/ + Mjnoi) on the vertical axis 
(Fig- El top right). In this plot, the sample galax- 
ies come closer to the universal relation. Only a 
few systems fall away from the overall relation be- 
cause they possess unusually high masses of gas. 



Their Mmoi values are apparently overestimated 
(see the next section). Thus, the total gas masses 
for most of the galaxies with anomalously high 
molecular-gas abundances are consistent with the 
angular momenta of the galaxies. Hence, the ex- 
cess H2 in such galaxies is not neither due to excess 
"portions" of molecular gas that have come to the 
galaxy disk, e.g., via the accretion of cold clouds, 
nor due to overestimates of the H2 abundances, 
but instead it is caused by the transition of most of 
the interstellar hydrogen into the molecular state. 

This conclusion remains unchanged if the radial 
photometric disk scale Rq is used instead of the 
optical radius (lower panel of Fig. O . Note that 
unlike the isophotal radius, Rq is not sensitive to 
differences between the surface brightnesses of the 
disks of the galaxies compared. 

3. POSSIBLE ORIGINS OF HIGH 
M^oi/M„i RATIOS 

Let us now list the main reasons that can lead 
to high estimates of the relative mass of molecular 
gas. 

1 . The actual conversion factors for some galax- 
ies may be much lower than the adopted 
value. 

2. This is the effect of the galaxy environment 
(accretion of large masses of molecular gas 
onto the disk, or the displacement inward of 
very cold, difficult to detect, molecular gas 
from the disk periphery due to interactions 
between galaxies 18]). 

3. An anomalously high content of dust, which 
serves as a catalyst of the formation of 
molecules and prevents the dissociation of 
molecules by the UV radiation of stars. 

4. Longer lifetimes of the gas in molecular form 
compared to most galaxies (e.g., a low star- 
formation rate per unit mass of molecular 
gas). 

5. Higher gas pressure in regions where it is 
mainly concentrated, which stimulates the 
transition of HI into H2- 

Bellow we analyze each of the above possibilities 
one-by-one as applied to our sample of galaxies 
with high Mmoi/Mni ratios. 

3.1. Conversion Factor 

The conversion factor X may be higher for 
galaxies with relatively high heavy-element abun- 
dances Metallicity of gas is usually associated 
with a high total luminosity L of a galaxy. Since 





Iog(a^|/<o^i>; 



-0.5 0.5 1 1.5 2 

logfo ,/<o ,>) 
mol mol ' 



FIG. 2: Histograms of HI and the molecular-gas surface density normalized to the mean values for the corre- 
sponding morphological types for the galaxies of the sample studied. 





11 



10 



•o.: 

"op 



go: 



2.5 



3.5 



log{V R ) 

rot o' 



11.5 
11 

10.5 
10 

i, 9-5 
-5 9 

8.5 
8 

7.5 























• 


• 




• 




< 


•*«*••• 




• 




t 


o 






• 


• 

Qd- 












o 










• 













2.5 



3.5 



log(V R ) 

rot o' 



FIG. 3: Mass of atomic gas (left) and total mass of gas (right) as a function of specific angular momentum defined 
as VrotR25 (top) or VrotRo (bottom) for galaxies of the sample considered (open circles) and for all galaxies of 
catalog T^l for which the corresponding estimates are available (filled circles) . R25 and _Ro are photometric radius 
and a radial scale length of a disk. 



the galaxies in our sample are not characterized by 
very high luminosities (Fig. ^ , this explanation is 
not appropriate for most of our galaxies. 

However, some of our sample galaxies may in- 
deed have anomalously low X values. Grounds 
for this possibility are provided by the fact that 
the estimated total gas mass Mgas (including he- 



lium) derived for the adopted X is unrealistically 
high — comparable to the total mass of the galaxy 
Mtot within its optical boundaries. It is obvious 
that Mtot must exceed the total mass of the disk 
(not to mention the mass of its gaseous compo- 
nent) if to take into account the masses of the 
dark halo and bulge. Figure ^ compares the total 



gas mass with indicative mass Mtot = K-ot^25/2G' 
(in solar units), which is approximately equal to 
a total mass of a galaxy within for all the 

CISM galaxies with available estimates of Vrot and 
D25- For galaxies located above the slanting line, 
Mgas > Mtot/ "2, and the gas masses are very likely 
to be overestimated. These galaxies make up a 
small fraction of all the galaxies considered, but 
about half of these objects belong to our sample 
of galaxies with a predominance of molecular gas 
(open circles). The same galaxies also exhibit a 
gas excess in the diagrams shown in Fig. |31 sup- 
porting the hypothesis that the conversion factors 
for these objects may have been strongly overesti- 
mated. Note that the exclusion of these galaxies 
does not alter the conclusions drawn above. 

Lower conversion factors may be associated not 
only with high metallicity of a gas, but also 
with the specific physical state of the interstel- 
lar medium. For example, the conversion factor 
should be low if molecular gas is concentrated not 
in virialized clouds, as it is usually assumed, but 
in a less dense diffuse medium ^j, which is quite 
possible in the case of the densest circumnuclear 
disks, where both clouds and the more tenuous in- 
tercloud gas are molecular. The example of M64, 
which dense molecular circumnuclear disk is ob- 
served in ^^CO and ^^CO lines, shows that diffuse 
molecular gas not only exists, but may even be 
responsible for the bulk of the CO-line luminos- 
ity, and the use of the standard conversion factor 
yields appreciably an overestimated mass of molec- 
ular gas 



left) compares the dust surface density defined as 
Mdust/ in our sample galaxies with the mean 
dust density for all CISM galaxies of the given mor- 
phological type. It is evident that galaxies with a 
predominance of molecular gas are, indeed, char- 
acterized by a higher mean dust content. How- 
ever, this is probably due solely to the higher mass 
of gas in these galaxies, since the distribution of 
Mdust /Mgas for our galaxies differs little from that 
for all the CISM galaxies (Fig.gl top right). Note 
that the Mdust estimate used here refers to so- 
called warm dust heated by stars, which emits pre- 
dominantly in the FIR (40 -^120 ^m). The mass 
of dust at very low temperatures, which weakly 
radiates in this spectral interval, cannot be taken 
into account rigorously. However, there are no 
grounds to believe that galaxies with a high con- 
tent of molecular gas detected in CO are also char- 
acterized by anomalous amounts of very cool dust. 
The location of the sample galaxies on a two-color 
diagram likewise does not indicate high color ex- 
cesses (Fig. El . The squares on the diagram show 
the mean color indices for galaxies of various mor- 
phological types [231 , which mark out the location 
of the normal color sequence. 

Observations show that the mass of dust in 
galaxies is most closely related to the mass of 
molecular gas . Figure [3 shows the Mdust — 
Mmoi diagram for CISM galaxies. The sample 
galaxies (open circles) lie on the common relation 
(lower panel in Fig. (S)), confirming that these ob- 
jects have a "normal" content of dust per unit mass 
of molecular gas (except for a few galaxies) . 



3.2. Effect of the Environment 

The effect of the environment is not the main 
factor determining the high Mmoi/Mni ratios for 
the galaxies in our sample. First, the sam- 
ple contains only a few strongly interacting sys- 
tems or galaxies in clusters (Table Sec- 
ond, with few exceptions, dense environments 
around galaxies are not associated with a sharp 
increase in Mmoi/MHi- According to de Mello 
et al. 0], single galaxies and galaxies located in 
more crowded environments differ only slightly in 
terms of Mmoi/Mni, and the mean Mmoi/Mni 
ratio for such galaxies remains less than unity. 
They report {\g{M^oi/ Mhi)) = -0.69 ± 0.59 
and —0.51 ± 0.46 for the the means and standard 
deviations for the high-density sample and the con- 
trol sample of isolated galaxies, respectively. 



3.3. High Dust Content 

Let us now compare the galaxies in terms of the 
mass of dust they contain, estimated from the far 
infrared (FIR) radiation of galaxies. Figure (top 



3.4. Long Lifetime of Gas in the Molecular 
State 

Molecular gas gives birth to stars whose radia- 
tion, in turn, leads to the dissociation of molecules. 
Therefore, the hypothesis that gas spends a long 
time in the molecular state resulting in the large 
Mmoi is equivalent to suggestion of a low star- 
formation rate per unit mass of molecular gas. It 
is worth noting that most of the sample galaxies 
with high Mmoi/Mni are not characterized by high 
star-formation rates, as is testified to by their lo- 
cation in the two-color diagram (Fig. O . Two ob- 
jects in the "blue" part of the diagram, marked 
by crosses, are interacting systems or "mergers" 
{MCG 5-32-20 and MCG 5-29-45), which appear 
to be undergoing bursts of star formation. Judging 
by their color indices, most of the galaxies are char- 
acterized by moderate or low star-formation rates. 
The only galaxy that falls away from the overall 
dependence due to its high (B-V) is NGC1482, 
which is a peculiar lenticular galaxy undergoing a 
burst of star formation in its central region, and is 
characterized by an extended region of bright line 
emission and gas outflow (superwind) (22i |. 



12r 



11.5 



11 



10.5 



10 ■ 



9.5 



9 



• o • • • o 



8.5 



7.5 L 



9.5 



10 



10.5 



11 



11.5 



12 



FIG. 4: Estimated total gas mass (including helium) Mgaa as a function of the total (indicative) mass of the 
galaxy Mtot, within the photometric radius i?25. Filled and open circles correspond to CISM galaxies and 
to galaxies in our sample, respectively. The dashed line corresponds to Mgas ~ Mtot/'2- The mass Mmoi is likely 
substantially overestimated for galaxies located above this line. 



Star formation is a multistage process. If star 
formation proceeds in a quasi-stationary mode, the 
mass of gas involved in each stage of the process 
is proportional to the duration of this stage. In 
this case, a large amount of molecular gas could be 
due to a relatively slow formation of gravitation- 
ally unstable regions inside molecular clouds (inef- 
ficient damping of turbulent motions?) or in the 
diffuse molecular gas, which evolves on a shorter 
time scale than the dense clouds. 

The current star-formation rate is closely re- 
lated to the infrared emission by dust Lpm, which 
absorbs the short- wavelength radiation of stars. 
To compare the star-formation efficiencies of the 
galaxies, the left and right panels of Fig. [T] show 
the relative mass of gas for the CISM galaxies with 
Lpm/Mgas and Lpm/M^oU respectively. The 
first dependence demonstrates an increase in the 
star- format ion efficiency with an increasing rela- 
tive H2 content: the interstellar gas is used up 
faster in galaxies that are rich in molecular hy- 
drogen than in those with small amounts of H2. 
However, this is true only for the total mass of 
gas, which in most cases is close to the mass of 
HI. However, as follows from the right-hand plot, 
the larger molecular fraction of gas, the lower, on 



the average, the star-formation rate per unit mass 
of molecular gas. Consequently, the intensity of 
the radiation is lower leading to molecular disso- 
ciation. This means that, on average, in galaxies 
dominated by H2 , the gas should spend more time 
in the molecular state before it becomes involved 
in the process of star-formation and dissociation. 

On both diagrams, two galaxies deviating 
strongly from the overall dependence show up con- 
spicuously in the top right corner are IC860 and 
NGC44I8. The anomalously high intensity of their 
infrared radiation is likely associated with bursts 
of star formation in these objects |2^, p^ . 



3.5. Excess Pressure of the Interstellar Gas 

The pressure of gas is determined mainly by two 
parameters — the gas surface brightness and the 
volume density of the disk, which is in good agree- 
ment with observations 0, |2^. A high gas 
pressure appears to be a crucial factor leading to 
the transformation of neutral (into molecular) hy- 
drogen 0]. 

A high pressure (density) of gas can be due 
both to the existence of local regions of com- 




FIG. 5: Dust content in galaxies. Top left: a histogram of the logarithm of the ratio of the surface density of dust 
to the mean value for the given morphological type according to the CISM catalog [inl |: top right: a histogram 
of the logarithm of the ratio of the mass of dust to the mass of gas (dark gray is for our sample of galaxies, light 
gray is for all galaxies of the CISM catalog with available Mmoi)', and at the bottom: the dependence of the mass 
of molecular gas on the mass of dust (filled circles are for all CISM galaxies, while open circles are for galaxies in 
our sample). 



pression (high contrast spiral arms, powerful star- 
forming regions associated with superclouds) and 
to a higher density of the stellar disk in the regions 
containing most of the gas. M51 is an example 
of a galaxy with a high relative content of molec- 
ular gas in its powerful spiral arms |23|. Giant 
gaseous complexes are concentrated in the spiral 
arms, and the gas velocity fields indicate compres- 
sion of the gas at the front of the shock associ- 
ated with them. In the absence of strong compres- 
sional waves, the gas would first be transformed 
into molecular form in regions with higher surface 
density and smaller thicknesses of the gaseous disk. 
Therefore, the highest gas pressures are expected 
where gas is concentrated in the inner region of 
the galaxy. Indeed, about half the gas-rich spiral 
galaxies (although these do not necessarily show a 
predominance of H2 over HI) with available high 
resolution data on the radial distribution of the 
CO intensity demonstrate a significant growth of 
the density of molecular gas toward the center in 
the inner-disk region |23, [2^ . 

Most of our sample galaxies with Mmoi/MHi > 
2 do not show such high-contrast and well-defined 



spiral structures as M51, but, judging from the 
Digital Sky Survey images, a circumnuclear region 
1 2 kpc in size can usually be distinguished by its 
high brightness, which is indicative of intense star 
formation in the dense central part of a disk. High 
pressures, and, consequently, a " molecularization" 
of the gas in the central region of the galaxy could 
be due to the high surface density of gas combined 
with the high density of the central part of the stel- 
lar disk, whose gravitational field compresses the 
gas into a thin layer, increasing its volume density. 

Unfortunately, data on the radial distribution of 
the gas are available for only a few of the galaxies 
in our sample. The best studied example is Ml 00, 
for which a high-resolution radial Ico profile was 
obtained with the BIMA ^ and Nobeyama Ob- 
servatory 01 interferometers. Most of the gas in 
this galaxy is in the molecular state, and is concen- 
trated in an inner region with a diameter of 2 3 
kpc (although the gas also remains mostly molec- 
ular at larger distances from the center). 

The radial motion of the gas toward the center 
could be due to both dynamical friction acting on 
giant molecular clouds and to the interaction of gas 



-0.6 
-0.4 
-0.2 
DO 
0.2 
0.4 
0.6 
0.8 



0.2 0.4 0.6 

B-V 



0.8 1 




(B-V) - <B-V> 



FIG. 6: Left: two-color diagram for the corrected color indices (U — B) — (B — V) (according to the HYPERLEDA 
database). Filled circles show the galaxies in our sample and squares are for the mean values for galaxies of various 
morphological types (according to [20|'). The X's denote interacting galaxies and the asterisks are for galaxies 
that belong to clusters. Right: distribution of the galaxies in our sample in the deviation of their B — V color 
indices from the mean values for the corresponding morphological type). 




1 

0.5 


J -0.5 
-1 

J 

I -1.5 
-2 
-2.5 
-3 











• 
• 


• 




• 






• • • 


• 




• 































0.3 1 



iog(M^^,/M^,) 



iog(M^^,/M^,) 



FIG. 7: Dependence of the star-formation efficiency defined as Ioq^Lfir/ Mgas) and as IoqILfir/ Mmoi) (right) on 
the relative content of molecular gas, log{AImoi/MHi), for CISM galaxies; the vertical line separates the sample 
galaxies with hight fraction of Mmoi- 



with a currently or previously existing bar (recall 
that more than half of our galaxies have apprecia- 
ble bars). 



4. CONCLUSIONS 

Galaxies whose gas is predominantly in molecu- 
lar form {Mmoi/MHi > 2) occur among disk galax- 
ies of all morphological types, although Sc — Sd 
types are among these objects. These galaxies dif- 
fer only slightly from galaxies with lower relative 
mass of molecular gas in terms of their luminosi- 
ties, surface brightnesses, rotational velocities, and 
dust contents per unit mass of gas. At the same 
time, galaxies that are rich in molecular gas show 
a tendency to have bars and active nuclei (to be 
LINERs) , providing indirect evidence for a concen- 
tration of gas toward the nucleus. 

The high Mmoi/Mni ratios in the galaxy sam- 
ple we have considered is not due to a loss of HI 



or the acquisition of H2, but instead is the result 
of the transformation of most of the HI into H2, 
since the total gas mass, Mgas, remains normal and 
consistent with the specific angular momentum of 
the disk. However, the Mmoi values for some of 
the sample galaxies derived from CO-line intensi- 
ties are probably appreciably overestimated due to 
the use of a conversion factor that appears to be 
overestimated. 

With only a few exceptions, galaxies with high 
Mrnoi/Mni ratios are characterized by moderate 
star-formation rates and comparatively high star- 
formation efficiencies, SFR/Mgas- However, the 
mean star-formation rate per unit mass of molecu- 
lar gas in these galaxies is appreciably lower com- 
pared to galaxies with low Mmoi/MHi- This means 
that, on average, the gas spends more time in 
molecular form before it becomes atomic again. 

As far as the integral properties of our sample 
galaxies are normal, their high content of molec- 
ular gas must be due to specific properties of the 



gas in local regions of the disk. The high frac- 
tion of molecular gas appears to be due to two 
interrelated factors. The first of these is a high 
density (pressure) of the gas due to its concentra- 
tion in the inner region of the disk (or in spiral 
arms, if the galaxy possesses a high-contrast spi- 
ral structure). In this case, the local dust den- 
sity should also be higher, facilitating the forma- 
tion and preservation of molecules. We think that 
there are no galaxies with a deficit of gas in their 
central region among those with a predominance 
of molecular gas. The second possible factor is the 
long duration of the molecular state of the gas dur- 
ing the chain HI — > H2 —>■ gravitationally unsta- 
ble clouds — > stars. Evidence for this is provided 
by the systematically lower star- format ion rate per 
unit mass of molecular gas for the sample galax- 
ies. If the gas is concentrated in the inner regions 
of the galaxy, this delay of its transformation into 
stars can be logically associated with the high an- 



gular velocity of rotation in the inner disk region, 
which enhances the gravitational stability of the 
gaseous disk against local perturbations, which, in 
turn, leads to the accumulation of gas and its sub- 
sequent transition into molecular form. 

We note, however, that the concentration of 
molecular gas toward the center of the galaxy is a 
fairly common (although according to [13 , it is less 
pronounced for galaxies of the latest morphologi- 
cal types). However, a predominance of molecular 
over atomic gas is a rare phenomenon, implying 
that a central concentration of a gas is essential, 
but no means a sufficient condition for such a pre- 
dominance. To understand better the causes for 
the transformation of most of HI into H2 on a 
galactic scale a more complete picture of gas evo- 
lution remains needed. 

The work was supported by RFBR grant 04-02- 
16518. 



[1] A. Boselli, J. Lequeux, and G. Gavazzi, As- 

tron. Astrophys. 384, 33 (2002); 
[2] N. Arimoto, Y. Sofue, and T. Tsujimoto, PASJ, 

48, 275 (1996); 
[3] D. Downes, P.M. Solomon, Astrophys. J. 507, 615 

(1998); 

[4] E. Rosolowsky, L. Blitz, Astrophys. J. 623, 826 
(2005); 

[5] B. G. Elmegreen, Astrophys. J. 338, 178 (1989); 
[6] B. G. Elmegreen, Astrophys. J. 411, 170, (1993); 
[7] T. Wong, L. Blitz, Astrophys. J. 569, 157 (2002); 
[8] C. F. McKee, J. P. Ostriker, Astrophys. J. 474, 
497 (1977); 

[9] F. Casoli, S. Sauty, M. Gerin, et al, Astron. As- 
trophys. 331, 451 (1998); 

[10] D. Bettoni, G. Galletta, and S. Garcia-Burillo, As- 
tron. Astrophys. 405,5 (2003); 

[11] HYPERLEDA http : //leda . univ-lyonl . f r/ 1 

[12] W. E. Baggett, S. M. Baggett, and K. S. J. An- 
derson, Astron. J. 116, 1626 (1998); 

[13] T. Boker, U. Lisenfeld, and E. Schinnerer, As- 
tron. Astrophys. 406, 87 (2003); 

[14] M.-P. Veron-Cetty, P. Veron, Astron. Astrophys. 
374, 92 (2001); 

[15] A. S. Evans, J. M. Mazzarella, et al., Astrophys. J. 
580, 749 (2002); 

[16] A. V. Zasov, A. A. Smirnova, AstL, 31, 160 
(2005); 

[17] I.D. Karachentsev, V.E. Karachentseva, 
W.K. Huchtmeier, et al., Astron. J. 127, 
2031 (2004); 

[18] D. Pfenniger, F. Combes and L. Martinet As- 
tron. Astrophys. 285, 79 (1994); 

[19] D.F. de Mello, T. Wiklind, and M.A.G. Mala, As- 
tron. Astrophys. 381, 771 (2002); 

[20] R. Buta, S. Mitra, G. de Vaucouleurs, et al., As- 
tron. J. 107, 118 (1994); 

[21] N. A. Devereux, J. S. Young, Astrophys. J. 359, 
42 (1990); 

[22] S. Veilleux, D. S.Rupke, Astrophys. J. 565L, 63 



(2002); 

[23] I. Kazes, H. Karoji, et al., Astron. Astrophys. 
197L, 22 (1988); 

[24] M. Imanishi, K. Nakanishi, N. Kuno, et al, As- 
tron. J. 128, 2037 (2004); 

[25] L. Blitz, E. Rosolowsky, Astrophys. J. 612L, 29 
(2004); 

[26] M. H. Heyer, E. Corbelli, et al, Astrophys. J. 602, 
723 (2004); 

[27] S. Aalto, S. Huttemeister, et al., Astrophys. J. 
522, 165 (1999); 

[28] M. W. Regan, M. D. Thornley, et al, Astro- 
phys. J. 561, 218 (2001); 

[29] K. Sakamoto, S. K. Okumura, et al., Astrophys. J. 
525, 691 (1999); 



TABLE 3: Galaxies with M^oi/Mhi > 2. 



n 


Name 




Type Dist. 


Ro 


L 


D25 




) lg(M«j 


) lg(M^„i 


;) lg(Kot 




(B - V)c Remarks 










Mpc 


Kpc 


Lq 


Kpc 


Mq 


Mq 


Mq 


km/s 








1 


NGC23 




1.2 


67 


0.57 


10.78 


1 


.56 


7.14 


9.95 


10.27 


2, 


.45 


— 


0.72 


B 


2 


NGC142 




3.0 


113 


— 


10.64 


1 


.53 


7.04 


9.73 


10.16 


2 


.10 


— 




B 


3 


JVGC695 




-2.0 


140 


— 


11.15 


1, 


.31 


7.64 


10.36 


10.81 


2, 


.45 


— 






4 


NGC828 




1.0 


78 


— 


10.88 


1, 


.82 


7.69 


10.15 


10.75 


2, 


.48 


0.32 


0.79 




5 


NGC972 




2.0 


23 


0.53 


10.35 


1, 


.34 


6.95 


9.13 


9.75 


2, 


.19 


0.07 


0.66 




6 


NGG1022 




1.1 


19 


0.26 


9.94 


1, 


.16 


6.13 


8.53 


9.24 


2, 


.02 


0.20 


0.69 


B 


7 


NGC1482 




-0.9 


24 


— 


9.72 


1 


.19 


6.58 


8.89 


9.68 


2 


.13 


-0.01 


0.86 




8 


UGC2982 




5.6 


75 


— 


10.75 


1, 


.25 


7.28 


10.14 


10.47 


2, 


.32 


— 




B 


9 


/C2040 




-1.0 


16 


— 


9.06 


0, 


.81 


5.04 


8.23 


8.73 


1, 


.74 


— 






10 


iVGC1819 




-2.0 


63 


— 


10.49 


1, 


.46 


6.91 


9.51 


10.13 


2, 


.25 


— 




B 


11 


iVGC2146 




2.3 


16 


— 


10.42 


1 


.42 


6.80 


9.67 


10.06 


2, 


.30 


0.18 


0.65 


B 


12 


NGC2775 




1.6 


19 


— 


10.46 


1 


.40 


6.89 


8.66 


9.49 


2 


.49 


0.32 


0.83 




13 


NGC3032 




-1.9 


24 


0.30 


9.66 


1 


.09 


5.79 


8.22 


8.91 


2 


.27 


0.09 


0.63 


B 


14 


NGC3147 




3.9 


44 


— 


11.02 


1 


.74 


7.51 


10.21 


10.60 


2 


.54 


— 


0.77 




15 


NGC3504, 




2.1 


24 


0.40 


10.30 


1, 


.26 


6.52 


8.78 


9.70 


2, 


.25 


-0.03 


0.67 


B 


16 


NGG3683 




5.0 


28 


0.18 


10.17 


1, 


.17 


6.76 


9.32 


9.66 


2, 


.27 


— 




B 


17 


NGC3758 




2.7 


129 


— 


10.63 


1 


,31 


6.89 


9.47 


9.90 


2 


.43 


— 






18 


NGC3860 




2.0 


81 


— 


10.48 


1 


.44 


7.01 


9.22 


9.54 


2 


.41 


0.25 


0.68 




19 


CGCG157 - - 


46 


3.0 


96 


— 


10.16 


1 


.38 


6.73 


9.03 


9.47 






— 






20 


AfGC3953 




4.0 


18 


0.59 


10.63 


1, 


.61 


7.26 


9.41 


9.96 


2, 


.35 


0.12 


0.68 


B 


21 


iVGC3993 




3.1 


71 


— 


10.54 


1, 


.53 


— 


9.80 


10.61 


2, 


.24 


— 






22 


iVGG4064 




1.4 


15 


— 


9.81 


1, 


.24 


5.54 


7.96 


8.48 


1, 


.90 


0.09 


0.67 


B 


23 


(7GC07064 




3.3 


109 


— 


10.62 


1 


.44 


7.63 


9.60 


10.21 


2, 


.42 


-0.02 


0.60 


B 


24 


7VGC4102 




3.0 


15 


— 


9.94 


1 


.14 


6.36 


8.75 


9.37 


2, 


.23 


— 




B 


25 


MCG5 - 29 - 


45 


3.2 


112 


— 


10.55 


1, 


.42 


6.74 


9.87 


10.78 


1, 


.97 


-0.51 


0.17 


B, VV 


26 


ArGC4245 




0.1 


15 


0.30 


9.65 


1, 


.12 


5.46 


7.80 


8.35 


2, 


.12 


0.43 


0.85 


B 


27 


iVGC4274 




1.7 


16 


0.63 


10.20 


1, 


.52 


6.32 


8.96 


9.31 


2, 


.37 


0.35 


0.83 


B 


28 


ArGC4298 




5.2 


18 


0.38 


10.10 


1, 


.18 


6.83 


8.99 


9.48 


2, 


.10 


— 


0.61 


V 


29 


7VGC4310 




-1.0 


15 


— 


9.05 


0, 


.97 


5.35 


7.88 


8.31 


1 


.90 


— 




B 


30 


ArGC4314 




1.0 


16 


-0.06 


10.11 


1 


.27 


5.67 


7.14 


8.98 


2, 


.33 


0.27 


0.81 


B 


31 


ArGC4321 




4.0 


24 


0.95 


11.04 


1, 


.72 


7.22 


9.93 


10.46 


2, 


.28 


-0.05 


0.65 


B, V 


32 


iVGG4394 




2.9 


15 


0.37 


9.91 


1, 


.17 


5.94 


8.56 


8.96 


2, 


.33 


0.19 


0.81 


B, V 


33 


iVGG4414 




5.1 


13 


0.23 


10.14 


1, 


.12 


6.50 


9.39 


9.70 


2, 


.35 


0.10 


0.77 


V 


34 


7VGC4418 




1.0 


31 


— 


9.69 


1 


.13 


6.31 


8.83 


9.34 


1 


.70 


— 






35 


WGC4419 




1.1 


22 


0.13 


10.31 


1 


.33 


6.39 


8.41 


9.83 


2, 


.05 


0.38 


0.81 


B, V 


36 


7VGC4448 




1.8 


12 


0.04 


9.54 


0, 


.94 


5.51 


7.81 


8.41 


2, 


.32 


0.33 


0.86 


B 


37 


iVGG4450 




2.3 


30 


0.81 


10.89 


1, 


.64 


7.08 


9.05 


9.39 


2, 


.30 


— 


0.76 


V 


38 


iVGG4457 




0.5 


13 


0.49 


9.76 


1, 


.03 


5.57 


8.32 


9.10 


2, 


.04 


0.23 


0.81 


B, V 


39 


JVGC4501 




3.3 


34 


0.89 


11.39 


1, 


.83 


7.68 


10.04 


10.59 


2, 


.46 


0.29 


0.63 


V 


40 


ArGG4548 




3.1 


8 


0.07 


10.59 


1 


.10 


6.14 


8.29 


8.76 


2, 


.29 


0.38 


0.75 


B 


41 


7VGC4579 




2.8 


23 


0.68 


10.82 


1 


.57 


6.86 


9.06 


9.86 


2, 


.44 


0.41 


0.75 


B 


42 


AfGC4580 




1.6 


16 


0.15 


9.55 


0, 


.96 


5.91 


7.66 


8.15 


2, 


.02 


— 




B 


43 


ArGC4689 




4.7 


25 


0.66 


10.44 


1, 


.51 


6.62 


9.09 


9.81 


2, 


.09 


— 


0.60 


V 


44 


iVGC4710 




-0.8 


20 


0.24 


10.13 


1, 


.46 


6.17 


8.06 


9.10 


2, 


.17 


0.29 


0.78 




45 


JVGG4818 




2.0 


15 


0.41 


10.02 


1 


.20 


5.87 


7.98 


9.12 


2, 


.12 


0.15 


0.76 


B 


46 


7VGC4848 




5.6 


104 


— 


10.83 


1 


.66 


6.96 


9.27 


9.58 


2, 


.08 


-0.14 


0.52 


B 


47 


ArGC4858 




3.0 


137 


— 


10.30 


1 


.31 


6.91 


9.05 


9.70 






-0.20 


0.36 


B, C 


48 


JC4040 




7.2 


115 


— 


10.45 


1, 


.44 


7.09 


9.07 


9.66 


1, 


.73 


-0.20 


0.39 


B, C 


49 


NGC4922 




-4.4 


104 


— 


10.73 


1, 


.58 


6.86 


9.35 


10.03 






0.54 


0.78 


VV, C 


50 


NGG4984 




-0.8 


17 


1.34 


9.86 


1, 


.18 


5.75 


8.24 


9.22 


2, 


.12 


0.28 


0.83 


B 


51 


/C860 




2.4 


58 


— 


10.03 


1 


.19 


6.76 


7.87 


9.03 


2, 


,37 


— 






52 


7VGC5054 




4.1 


24 


— 


10.51 


1 


.53 


6.73 


9.58 


9.96 


2, 


,27 


0.08 


0.63 




53 


UGC8399 




3.0 


107 


— 


10.37 


1 


.43 


6.57 


9.74 


10.62 


1 


,76 


— 




B 


54 


MCG5 - 32 - 


20 


8.0 


103 


— 


10.30 


1, 


.25 


5.96 


9.51 


10.13 


2, 


,26 


-0.40 


0.32 


VV 


55 


iVGG5187 




3.0 


105 


— 


10.36 


1, 


.49 


6.97 


9.99 


10.57 


2, 


,26 


— 






56 


ArGG5653 




3.0 


54 


— 


10.62 


1 


.43 


7.11 


9.69 


10.08 


2, 


,38 


-0.11 


0.62 




57 


ArGG5678 




3.3 


31 


— 


10.64 


1 


.48 


7.04 


9.53 


10.13 


2, 


,30 


— 




B 


58 


ArGC5676 




4.7 


34 


0.65 


10.79 


1 


.58 


7.18 


9.94 


10.26 


2, 


,39 


-0.01 


0.57 




59 


AfGG5936 




3.2 


59 


0.57 


10.59 


1 


.36 


7.07 


9.63 


10.12 


2, 


,46 


-0.12 


0.52 


B 


60 


ArGC6000 




4.0 


30 




10.31 


1, 


.22 


6.95 


9.82 


10.13 


2, 


,27 






B 


61 


ArGC6240 




0.0 


105 




10.92 


1, 


.80 


7.46 


10.01 


10.69 


2, 


,30 


0.25 


0.69 


VV 


62 


iVGG6574 




3.9 


35 




10.55 


1 


.17 


6.86 


9.24 


10.09 


2, 


,40 


0.02 


0.63 


B 


63 


AfGC6951 




3.9 


24 


0.75 


10.89 


1 


.39 


6.92 


9.72 


10.11 


2, 


,35 




0.60 


B 


64 


JVGG7225 




-0.3 


68 




10.51 


1 


.59 


7.03 


9.38 


10.07 


2, 


,21 








65 


NGG7770 




0.0 


62 




10.17 


1, 


.18 




9.81 


10.49 


2, 


,59 




0.47